ABSTRACT
The degree of componentiality assigned to phonological representations has been the subject of some debate within recent years, and various proposals have been made for subdividing the segment into clearly defined subgroupings, or GESTURES. In autosegmental phonology, features occupy different autosegmental tiers according to the nature of the particular process in question. Thus, in a process of nasal spreading, the feature [nasal] might occupy an autosegmental tier whose boundaries do not coincide with those of segments on the other tier(s), while in other processes a different feature may be extracted from the whole, and the utterance auto-segmentalised in a different way. The notational systems of metrical phonology and dependency phonology start out from the same fundamental belief with respect to suprasegmental structure — that we need to set up a non-linear system of representation to capture relative prominence holding both between segments and between suprasegmental units.
